On Thu, Jan 15, 2004 at 07:44:09PM +0100, Leonard den Ottolander wrote: > Hi Charles, > > > > -rwxr-xr-x 1 root root 83684 Sep 5 21:14 /sbin/ethtool > > > -rwxr-xr-x 1 root root 83684 Sep 5 14:14 /sbin/ethtool > > > > > > Notice that while my length and date agree with yours, my time and > > md5sum do not. > > I explained the md5sum mismatch in my other post (indeed prelink). You did, and I ran your command line on three of the four machines (after forcing prelink on the two I had just updated). They agree: [root@issola root]# /usr/sbin/prelink -N --verify --md5 /sbin/ethtool 664b71f93f11aac80957f19273288f01 /sbin/ethtool One machine of the four does not have prelink on it, and it gave the same md5sum. [root@jhereg root]# md5sum `which ethtool` 664b71f93f11aac80957f19273288f01 /sbin/ethtool > The difference in date is because your RTC probably uses local time, > and his GMT. Good point. > > > Question: is prelinking the culprit on the length and md5sum > > differences? > > I haven't read anything on length mismatches. OK. I note that the description of prelink in "rpm -qif /usr/sbin/prelink" indicates that it modifies both libraries and executables. It does not indicate whether prelink changes the length. If the different lengths give different md5sums, but when you run the md5sum through prelink and get the same md5sum, I conjecture that prelink accounts for the length difference. Also, if I examine the ctime, it shows the time when I last ran prelink. E.g: [root@issola root]# ll -c `which ethtool` -rwxr-xr-x 1 root root 83456 Jan 15 12:15 /sbin/ethtool On the machine which does not have prelink, the ctime is the installation date. -- Charles Curley /"\ ASCII Ribbon Campaign Looking for fine software \ / Respect for open standards and/or writing? X No HTML/RTF in email http://www.charlescurley.com / \ No M$ Word docs in email Key fingerprint = CE5C 6645 A45A 64E4 94C0 809C FFF6 4C48 4ECD DFDB
Attachment:
pgpDgj1utlO03.pgp
Description: PGP signature
